More people in the U.S. die from lung cancer than from prostate, breast, colon and rectum cancers combined. Of the estimated 187,000 people in the U.S. who were diagnosed with non-small cell lung cancer (NSCLC) in 2009, over 37,000 of them presented with localized or early stage disease. This fraction is expected to grow as methods of early detection improve and become more widely disseminated. Stereotactic body radiation therapy (SBRT) has been shown to provide excellent local control (85%-95%) for early-stage lung cancer patients. However, a recent phase II study found Grade 3 or greater toxicities in a significant fraction of the patients, particularly those with centrally located tumors. Studies have shown that decreases in SBRT margins can significantly reduce the probability of normal tissue complications. As SBRT is increasingly becoming the therapy of choice for early-stage, localized non-small cell lung cancer, reducing the harmful side effects becomes increasingly important. We are proposing a novel failsafe technology to reduce toxicity while retaining local control for this growing population of patients. Our hypothesis is that tracking lung tumors directly during SBRT, using beam's-eye-view (BEV) imaging coupled with a dynamic multileaf collimator (DMLC), will lead to clinically significant normal tissue sparing. The current proposal is the first to employ an advanced multi- template marker-less BEV tracking algorithm to derive the real-time tumor location for DMLC delivery. Clinical benefits of the multi-template marker-less innovation include 1) automatic selection of relevant landmarks, 2) continuous tracking during deformations, rotations and partial obscurations, 3) no additional imaging dose to the patient, 4) direct imaging of the entire tumor, and 5) no need for invasive fiducial implantations and the associated adverse effects. This represents a substantial improvement over previous techniques. The therapeutic advantage will be quantified experimentally in an anthropomorphic phantom system under clinical conditions. The end result will be an integrated real-time target tracking and dynamic delivery system for radiation therapy as well as quantification of the anticipated clinical benefits. Clinical integration with the DMLC tracking is a challenging engineering problem with a measurable positive impact on human health on a large scale. This project represents well the ideals of the NIBIB to support research and development in the physical sciences and engineering for the improvement of human health. PUBLIC HEALTH RELEVANCE: The goal of this project is to quantify the potential clinical benefits of tracking early stage lung tumors with the therapeutic radiation beam in order to avoid over irradiation of healthy tissues during treatment. Our proposed method is the safest, least intrusive to have been proposed. If found to be successful, the technique could be applied to other anatomical sites, providing better outcomes for a very large number of cancer patients.